Microbiology (2004), 150, 1103–1111
DOI 10.1099/mic.0.26695-0
Effect of pyruvate kinase overproduction on
glucose metabolism of Lactococcus lactis
Ana Ramos,1 Ana Rute Neves,1 Rita Ventura,1 Christopher Maycock,1
Paloma López2 and Helena Santos1
1
Instituto de Tecnologia Quı́mica e Biológica/Universidade Nova de Lisboa and Instituto de
Biologia Experimental e Tecnológica, Rua da Quinta Grande, 6, Apt 127, 2780-156 Oeiras,
Portugal
Correspondence
Helena Santos
santos@itqb.unl.pt
2
Centro de Investigaciones Biológicas, Velazquez 144, Madrid, Spain
Received 8 August 2003
Revised 19 December 2003
Accepted 5 January 2004
Lactococcus lactis strain NZ9000(pNZpyk), which overproduces pyruvate kinase (PK),
was constructed. The pNZpyk plasmid carries the PnisA–pyk transcriptional fusion, and the
overexpression of its pyk gene was accomplished by using the nisin-inducible expression system
of the NZ9000 strain. In vivo 13C- and 31P-NMR spectroscopy was used to evaluate the effect
of this modification on the metabolism of glucose in non-growing cells. A detailed description
of the kinetics of glucose, end products, glycolytic intermediates, NAD+ and NADH was obtained.
A 15-fold increase in the level of PK did not increase the overall glycolytic flux, which, on the
contrary, was slightly reduced. Significant differences were observed in (i) the level of
3-phosphoglycerate (3-PGA) and phosphoenolpyruvate (PEP), metabolites associated with
starvation; (ii) the rate of fructose 1,6-bisphosphate (FBP) depletion upon glucose exhaustion;
and (iii) the NAD+/NADH ratio during glucose catabolism. In the mutant, the rate of FBP
consumption after glucose depletion was notably accelerated under anaerobic conditions, whereas
3-PGA and PEP decreased to undetectable levels. Furthermore, the level of NAD+ decreased
steadily during the utilization of glucose, probably due to the unanticipated reduction in the lactate
dehydrogenase activity in comparison with the control strain, NZ9000(pNZ8020). The results
show that PK is an important bottleneck to carbon flux only when glucose becomes limiting;
in the overproducer this constriction was no longer present, as evidenced by the faster FBP
consumption and lack of accumulation of 3-PGA and PEP in anaerobic as well as aerobic
conditions. Despite these clear changes, the PK-overproducing strain showed typical homolactic
metabolism under anaerobic conditions, as did the strain harbouring the vector plasmid without
the pyk insert. However, under an oxygen atmosphere, there was increased channelling of
carbon to the production of acetate and acetoin, to the detriment of lactate production.
INTRODUCTION
Lactic acid bacteria are suitable model systems for studying
the mechanisms involved in the regulation of sugar transport and metabolism. In particular, the relative metabolic
simplicity of Lactococcus lactis makes it an ideal candidate
for these studies. Regrettably, in spite of the wealth of
metabolic information and ingenious genetic tools available
for L. lactis, we are still far from having a comprehensive
understanding of sugar metabolism and regulatory pathways in this model organism. The accomplishment of this
Abbreviations: FBP, fructose 1,6-bisphosphate; G6P, glucose 6phosphate; GAPDH, glyceraldehyde-3-phosphate dehydrogenase (EC
1.2.1.12); KPi, potassium phosphate; LDH, L-lactate dehydrogenase
(EC 1.1.1.27); Pi, inorganic phosphate; PEP, phosphoenolpyruvate; PFK,
phosphofructokinase; PK, pyruvate kinase (EC 2.7.1.40); 2-PGA, 2phosphoglycerate; 3-PGA, 3-phosphoglycerate.
0002-6695 G 2004 SGM
Printed in Great Britain
goal has been hampered by the complexity of global metabolism, which consists of multiple interlocked pathways
connected via common metabolites and cofactors through
various degrees of genetic and metabolic regulation (Bailey,
2001). Molecular biology has been useful in dissecting complex systems into parts that are easier to analyse and understand; however, for obtaining a deep insight into metabolic
regulation, a comprehensive knowledge of the cells as a whole
is required. This achievement will be essential for the rational
design of reliable engineering strategies to reroute carbon
fluxes to the production of desired products.
Glycolytic regulation in L. lactis has been related to the high
fructose 1,6-bisphosphate (FBP) concentration that activates lactate dehydrogenase (LDH) and pyruvate kinase
(PK), directing the flux towards the production of lactate.
Accordingly, accumulation of phosphoenolpyruvate (PEP)
and 3-phosphoglycerate (3-PGA) in starved cells is regarded
1103
A. Ramos and others
as a consequence of PK inhibition by high inorganic phosphate (Pi) and low FBP concentrations associated with sugar
depletion (Mason et al., 1981; Thompson, 1987). Recent
studies have questioned the role of FBP and postulated that
inhibition or activation exerted by the cellular NADH/
NAD+ ratio on glyceraldehyde-3-phosphate dehydrogenase
(GAPDH) or LDH is the main issue regulating glycolysis
and the shift from homo- to heterolactic fermentation
(Garrigues et al., 1997). On the other hand, based on noninvasive NMR measurements of NAD+ and NADH, we
have shown that the glycolytic flux in resting cells is not
appreciably influenced by the NADH/NAD+ ratio (Neves
et al., 2002b). Recently, using an approach that involves
manipulation of enzyme activities in small steps above and
below the wild-type level and metabolic control analysis,
Jensen and collaborators have shown that GAPDH has no
control over glycolytic flux (Solem et al., 2003). Similarly,
LDH and also phosphofructokinase (PFK) have been ruled
out as major sites of control of lactate production in L. lactis
(Andersen et al., 2001b; Koebmann et al., 2002b). Hence, the
control of glycolysis in L. lactis remains elusive.
PK (ATP : pyruvate 2-O-phosphotransferase, EC 2.7.1.40)
catalyses the last step of glycolysis, where the phosphoryl
group of PEP is transferred to ADP to form pyruvate and
ATP. The reaction is irreversible under physiological conditions and has long been considered as critical for the
regulation of metabolic flux in the second part of glycolysis.
This central position is evident from the regulatory properties of PK, a typical allosteric protein (Valentini et al.,
2000). The lactococcal enzyme has been purified and characterized (Crow & Pritchard, 1976, 1982). It is activated by
FBP, the activity being a sigmoidal function of the activator
concentration, and the affinity for PEP and ADP is also
enhanced by FBP. Furthermore, the Ka for FBP increases
drastically in the presence of low concentrations of Pi. The
magnitude of the FBP and Pi pools in L. lactis has been
assessed in several studies; FBP accumulates up to 50 mM in
the presence of glucose or lactose, whereas high concentrations of Pi, 3-PGA, 2-PGA and PEP are typical of starved
cells (Thompson & Thomas, 1977; Thompson & Torchia,
1984; Neves et al., 1999, 2002b).
Early studies provided evidence for a pivotal regulatory role
of PK in the glycolytic metabolism of L. lactis (Mason et al.,
1981; Thompson & Torchia, 1984). These data prompted us
to evaluate the effect of PK overproduction on the physiology
of this model organism and to investigate the role of this
enzyme in the regulation of the overall glycolytic flux. NMR
techniques were used to monitor the kinetics of the intracellular pools of several glycolytic metabolites, NAD+, NADH,
ATP and Pi in living cells of L. lactis overproducing PK.
METHODS
Materials. Restriction enzymes, T4 DNA ligase, alkaline phospha-
tase, FBP, PEP, LDH and NADH were purchased from Roche
Molecular Biochemicals. DyNAzyme DNA polymerase was from
Finnzymes. Oligonucleotides were synthesized by MWG-Biotech.
1104
[1-13C]Glucose (99 % isotopic enrichment) and [6-13C]glucose (99 %)
were supplied by Campro Scientific. Other chemicals were reagent
grade and obtained from Sigma or Aldrich.
Bacterial strains and media. Escherichia coli MC1061 was grown
in Luria–Bertani broth with aeration at 37 uC. Lactococcus lactis
strains MG1363 (Gasson, 1983), NZ9000 (expressing the nisRK regulatory genes; Kuipers et al., 1998) and its derivatives were routinely
cultivated without aeration in M17 medium supplemented with 1 %
(w/v) glucose. Chloramphenicol was used at a concentration of 20
or 5 mg ml21 for selection of pNZ8020-based plasmids (de Ruyter
et al., 1996) in E. coli or L. lactis, respectively. Bacterial growth was
monitored spectrophotometrically at 600 nm.
DNA techniques and transformation. All manipulations with
recombinant DNA in E. coli MC1061 (used as an intermediate host
for cloning) were carried out according to standard procedures
(Sambrook et al., 1989) and as specified by the enzyme manufacturers.
Chromosomal and plasmid DNA of L. lactis was isolated as described
by Leenhouts et al. (1989) and O’Sullivan & Klaenhammer (1993),
respectively. PCR reactions were performed in a total volume of 100 ml
with 10 mM Tris/HCl (pH 8?8), 2 mM MgCl2, 50 mM NaCl, 200 mM
each deoxyribonucleoside triphosphate, 1 U DyNAzyme, 50 pmol of
each primer and 1 mg template DNA obtained from MG1363.
Amplification was performed in 25 cycles, each cycle consisting of a
denaturation step at 95 uC for 1 min, a primer annealing step at 55 uC
for 1 min, and a primer extension step at 72 uC for 2 min. Preparation
of competent L. lactis cells and electrotransformation was performed
as described by Dornan & Collins (1990).
Construction of plasmids and mutant strains. Primers Pyk1
(59-GCGCTGGATCCCGCATAATTAGGGG-39) and Pyk2 (59-CGTATGAATTCGCCTGTCGATAGCGG-39), based on the sequence of
the PK gene (accession no. LO7920) published previously (Llanos
et al., 1993), were designed to introduce BamHI and EcoRI restriction sites (underlined in the primer sequences) upstream and downstream, respectively, of the pyk gene coding region. A PCR product
with the expected size (1?6 kb) was obtained and cloned as a BamHI–
EcoRI fragment into the similarly digested, high-copy-number shuttle vector pNZ8020 under control of the nisA promoter. The ligation
mixture was transformed into E. coli MC1061, and a recombinant
plasmid carrying the appropriate insert (designated pNZpyk) was
selected. The pNZpyk and pNZ8020 plasmids were independently
introduced by electroporation to L. lactis NZ9000 for further studies.
Fermentation under controlled conditions. L. lactis cultures of
NZ9000(pNZpyk) and NZ9000(pNZ8020) were grown at 30 uC in a
2 litre fermenter in the defined medium (CDM) described by Poolman
& Konings (1988), containing 1 % (w/v) glucose, 5 mg chloramphenicol ml21 and 1 ng nisin ml21. Strain NZ9000 was grown in the
same medium without addition of chloramphenicol and nisin. For
the measurements of pyridine nucleotide pools, [5-13C]nicotinic acid
(5 mg l21), synthesized as described before (Neves et al., 2002a),
was supplied, and aspartate and asparagine were omitted. The pH
was kept at 6?5 by automatic addition of 10 M NaOH and a low
agitation rate (70 r.p.m.) was used to keep the fermentation broth
homogeneous. Anaerobic conditions were established by gassing the
medium with argon for 2 h before inoculation (4 % inoculum with
a culture grown overnight in the same medium); for aerobic growth,
the dissolved oxygen was monitored with an oxygen electrode
(Ingold). A specific air tension of 90 % was maintained by automatic
control of the airflow. Growth was evaluated by measuring the turbidity of the samples at 600 nm (OD600) and calibration against cell
dry weight measurements.
Preparation of ethanol extracts for analysis by 31P-NMR and
quantification of phosphorylated metabolites. Ethanol extracts
were prepared from culture samples (approx. 70 mg protein for each
Microbiology 150
Overproduction of pyruvate kinase in L. lactis
extract) obtained during growth as previously described (Ramos
et al., 2001), and the concentration of intracellular metabolites was
calculated from the peak areas in 31P spectra by comparing with the
area of the resonance due to methylphosphonic acid added as internal standard. Assignment of resonances was achieved by addition of
the pure compounds to the extracts or by comparison with previous
studies (Ramos & Santos, 1996; Ramos et al., 2001).
In vivo NMR experiments. Cells were harvested in the midexponential phase (OD600 2?2), centrifuged and suspended in 50 mM
potassium phosphate (KPi) buffer, pH 6?5, to a protein concentration of approximately 13 mg ml21. In vivo NMR experiments were
performed under anaerobic (argon atmosphere) or aerobic conditions (oxygen saturation) as described previously (Neves et al.,
1999). [1-13C]Glucose (or [6-13C]glucose) was added and the timecourse for its consumption, product formation, and build-up and
consumption of intracellular metabolite pools was monitored. After
glucose exhaustion, and when no changes in the resonances due to
end products and intracellular metabolites were observed, a cell
extract was prepared (Neves et al., 1999): briefly, an aliquot of the
cell suspension was passed through a French press (twice at 120 MPa);
the resulting extract was incubated at 80 uC for 15 min in a sealed
tube, cooled on ice; the cell debris was removed by centrifugation
(30 000 g, 10 min, 4 uC) and the supernatant (denoted hereafter as
total NMR-sample extract) was stored at 220 uC.
An aliquot of the cell suspension used for in vivo NMR was saved to
prepare a crude extract and measure the level of PK overproduction.
Protein concentration in cell suspensions was determined by the Lowry
method using bovine serum albumin as a standard after disruption of
cells by treatment with 1 M NaOH and incubation at 80 uC for 5 min.
Quantification of end products. Lactate, acetate and formate were
quantified in total NMR sample extracts by 1H-NMR as described
elsewhere (Neves et al., 1999). The concentrations of metabolic intermediates and minor end products that remained inside the cells (PEP
and 3-PGA) were determined by 13C-NMR. The concentration of
labelled lactate determined by 1H-NMR was used as a standard to
calculate the concentration of the other metabolites in the sample.
Carbon balances from glucose were always above 95 % after taking
into account both extracellular and intracellular metabolites.
Quantification of intracellular metabolites in living cells by
13
C-NMR. Due to the fast pulsing conditions used for acquiring
in vivo 13C-spectra, there is no direct correlation between concentrations and peak intensities. The correction factors that allowed the
conversion of peak areas of glucose 6-phosphate (G6P), FBP, 3PGA, NAD+ and NADH into intracellular concentrations were
determined as described previously (Neves et al., 1999, 2002a). The
quantitative kinetic data for intracellular metabolites were calculated
from their peak areas by applying the correction factors and comparing with the intensity of the lactate resonance in the last spectrum of the sequence. Intracellular metabolite concentrations were
calculated using a value of 2?9 ml (mg protein)21 for the intracellular volume (Poolman et al., 1987).
NMR spectroscopy. All NMR experiments with living cells were
run at 30 uC with a quadruple-nucleus probe head on a Bruker
DRX500 spectrometer as described before (Neves et al., 1999, 2002b).
Although individual experiments are illustrated in each figure, each
type of in vivo NMR experiment was repeated at least twice and the
results were highly reproducible. The values reported are means of
two or three experiments and the accuracy varied from 2 % (for end
products) to 15 % (for intracellular metabolites with concentrations
below 5 mM). The concentration limit for detection of intracellular
metabolites in living cells was around 0?6 mM in the experiments
http://mic.sgmjournals.org
where the pools of nicotinamide nucleotides were monitored (2?2 min
acquisition time for each spectrum), and 3–4 mM in the remaining
experiments (30 s acquisition).
Quantification of the pools of phosphorylated metabolites in growing
cells and measurements of PK activity by 31P-NMR were performed as
described previously (Ramos et al., 2001). Carbon and phosphorus
chemical shifts were referenced to the resonances of external methanol
or 85 % H3PO4 designated at 49?3 p.p.m. and 0?0 p.p.m., respectively.
Preparation of crude cell extracts for determination of enzyme
activities. Cells were cultivated in defined medium as described
above until mid-exponential growth phase, harvested, washed and
suspended in 50 mM MOPS buffer pH 6?5. Cell extracts were prepared by mechanical disruption using a French press (three passages
at 120 MPa), debris was removed by centrifugation (30 000 g,
20 min, 4 uC), and the extracts were used immediately for determination of enzyme activity. The protein concentration in crude cell
extracts was measured by the method of Bradford (1976) using
bovine serum albumin as a standard.
Purification of PK. PK was purified to electrophoretic homogeneity
from approximately 65 g (wet wt) of cells of the NZ9000(pNZpyk)
strain essentially as described by Crow & Pritchard (1976).
Enzyme assays. PK was routinely assayed in cell extracts by the
spectrophotometric method described by Crow & Pritchard (1976).
LDH, NADH oxidase (Garrigues et al., 1997) and PFK (Fordyce
et al., 1982) were measured as described before.
The effect of activators and inhibitors of PK was determined by 31PNMR spectroscopy using the purified enzyme. The 3 ml reaction
mixtures contained 100 mM Tris/HCl buffer pH 7?5, 5 mM MgCl2,
10 mM KCl, 5 mM FBP, 2 mM ADP and 50 mg pure enzyme. The
reactions were started by addition of 2 mM PEP and the time-course
for its consumption was monitored; the effect of Pi, FBP, NADH,
NAD+, 3-PGA and ATP on the activity of PK was studied by adding
these compounds to the reaction mixtures. The reactions were stopped
before PEP exhaustion by incubation at 80 uC for 5 min and the rate of
PEP consumption was calculated by comparison of the area of the PEP
resonance to that of a known amount of methylphosphonate added as
an internal standard. One unit of enzyme activity was defined as the
amount of enzyme catalysing the conversion of 1 mmol PEP min21
under the experimental conditions used.
All the determinations were made at least in triplicate in two extracts
obtained from independent cultures.
RESULTS
Overexpression of the lactococcal PK gene
( pyk)
The pyk gene of L. lactis MG1363 was cloned under the
control of a nisin-inducible promoter in plasmid pNZpyk,
and overexpressed in L. lactis NZ9000 by addition of nisin
during growth (see details in Methods). In chemically defined
medium, a 15-fold higher level of active PK [27?6 units (mg
protein)21] was obtained in strain NZ9000(pNZpyk) after
induction with nisin (1 ng ml21), whereas the activity
found in NZ9000 harbouring the vector pNZ8020 was
similar to that detected in the model organism MG1363
[1?8 units (mg protein)21].
1105
A. Ramos and others
Pools of glycolytic intermediates monitored by
in vivo NMR
50
40
60
50
30
40
20
30
20
FBP (mM)
Glucose, Lactate (mM)
70
10
10
(a)
10
20
30
40 50 60
Time (min)
70
80
90
80
50
40
60
50
30
40
20
30
20
FBP, 3-PGA (mM)
Glucose, Lactate (mM)
70
10
10
(b)
10
20
30
40
50
60
70
80
90
Time (min)
7.4
45
40
35
30
25
20
15
10
5
7.2
7.0
6.8
6.6
6.4
Cytoplasmic pH
ATP, Pi (mM)
Strains NZ9000(pNZpyk) and NZ9000(pNZ8020) were
grown in defined medium containing 1 ng nisin ml21,
harvested in the mid-exponential phase, and glucose metabolism was studied in vivo by NMR. Lactate production, glucose
consumption and glycolytic intermediates during the metabolism of [1-13C]glucose (40 mM) in both strains under an
argon atmosphere are shown in Fig. 1(a, b). The conversion of
glucose was homofermentative in strains NZ9000(pNZpyk)
and NZ9000(pNZ8020), the glucose consumption rates being
0?19±0?01 and 0?22±0?02 mmol min21 (mg protein)21,
respectively. The reduction of the glucose consumption rate
in NZ9000(pNZpyk) when compared to the reference strain
NZ9000(pNZ8020) is noteworthy; a similar trend was
observed in a L. lactis construct overproducing NADH oxidase under the control of the nisin promoter (Neves et al.,
2002a). A striking difference in the profile of depletion of the
FBP pool was found. Although the maximal intracellular
concentration reached (approx. 45 mM) was similar in both
strains examined, the consumption of FBP after glucose
exhaustion proceeded at a much higher rate in the PKoverproducing strain. In addition, 3-PGA accumulated to a
maximal concentration of 10 mM in NZ9000(pNZ8020),
whereas in NZ9000(pNZpyk) this metabolite was always
below the in vivo detection limit (Fig. 1a, b). It is worth
mentioning that 3-PGA was also not visible in the 13C-NMR
spectrum of the NMR-sample extract (not shown) obtained
as described in Methods, meaning that its intracellular concentration was below 1?8 mM (detection limit). This result
was due to the overproduction of PK in NZ9000(pNZpyk)
since this strain behaved like NZ9000(pNZ8020) with respect to the pattern of FBP and 3-PGA accumulation when
intermediate levels of PK (7 units mg21) were achieved by
lowering the concentration of nisin to 0?25 ng ml21 (data
not shown).
80
6.2
ATP and intracellular Pi concentrations as well as the
evolution of intracellular pH during the anaerobic metabolism of glucose by NZ9000(pNZpyk) were determined by
31
P-NMR under conditions similar to those used for 13CNMR experiments (Fig. 1c). Upon substrate addition, the
concentration of ATP increased from undetectable levels
to a maximum of 6 mM. After glucose exhaustion (at
t<15 min) a sudden rise of intracellular Pi to about 25 mM
was followed by a gradual increase in concentration, reaching values in the order of 40 mM, similar to those of cells in
the initial stage of the experiment. The intracellular pH rose
from 6?2 in starved cells to 7?3 during active glycolysis and
decreased slowly after substrate depletion. These results
are similar to those obtained for strain MG1363 under
equivalent experimental conditions (Neves et al., 2002b).
Fig. 1. (a, b) Metabolism of glucose under anaerobic conditions
at 30 6C, pH 6?5, assessed by 13C-NMR in vivo: time-course for
the consumption of [1-13C]glucose (40 mM), lactate formation
and evolution of the FBP pool in (a) NZ9000(pNZpyk) and (b)
NZ9000(pNZ8020). (c) Intracellular pH, KPi and ATP concentrations were determined in the PK-overproducer by 31P-NMR after a
pulse of glucose using the same experimental conditions. Glucose
was added at time 0 and the shaded area indicates glucose availability. X, Glucose; %, lactate; m, FBP; #, 3-PGA; ,, ATP;
., Pi; 6, cytoplasmic pH.
Fig. 2 shows the data obtained during the metabolism of
[6-13C]glucose (40 mM) by both strains under aerobic
(oxygen-saturated) atmosphere. The use of glucose labelled
on C-6 allowed us to detect and quantify the resonance of
G6P, which otherwise was obscured by the intense peaks due
to [1-13C]glucose. The rates of glucose consumption by
NZ9000(pNZpyk) and NZ9000(pNZ8020) were 0?12±0?01
and 0?16±0?01 mmol min21 (mg protein)21, respectively.
Lowered glycolytic fluxes have been consistently observed
1106
(c)
6.0
_10 0 10 20 30 40 50 60 70 80 90
Time (min)
Microbiology 150
Overproduction of pyruvate kinase in L. lactis
Fig. 2. Effect of oxygen on the metabolism of glucose monitored by in vivo 13C-NMR: time-course for the consumption of
[6-13C]glucose (40 mM) and end product formation (a, b) and evolution of the intracellular metabolite pools (c, d) in strains
NZ9000(pNZpyk) (a, c) and NZ9000(pNZ8020) (b, d). X, Glucose; %, lactate; n, acetate; $, acetoin; m, FBP; e, G6P;
#, 3-PGA; &, PEP.
under aerobic conditions in several L. lactis strains (Neves
et al., 2002a). The behaviour of strain NZ9000(pNZ8020)
(Fig. 2b, d) was comparable to that reported for MG1363
under an oxygen atmosphere (Neves et al., 2002b): after the
addition of glucose, FBP and G6P rose to approximately 48
and 5 mM, respectively, decreasing to undetectable levels
after glucose depletion, whilst 3-PGA and PEP accumulated
to concentrations of approximately 32 and 10 mM, respectively, levels that are much higher than those measured
under an argon atmosphere. Acetate (7?8 mM) and acetoin
(1?1 mM) were produced in addition to lactate (67?2 mM).
In the NZ9000(pNZpyk) strain (Fig. 2a, c), the production
of acetate and acetoin was, respectively, two- and eightfold
higher than in NZ9000(pNZ8020). 3-PGA and PEP were
not detected, either in vivo or in NMR-sample extracts. The
acceleration of the FBP consumption rate occurred in both
strains upon switching from anaerobic to aerobic conditions
(compare Figs 1 and 2), but the maximal FBP level (approx.
37 mM) in the NZ9000(pNZpyk) strain was clearly lower
under an oxygen atmosphere as compared to anaerobic
conditions (approx. 45 mM).
of strain MG1363 and an LDH-deficient strain (Neves et al.,
2002b). A similar approach was followed here to assess
the magnitude of the NAD+ and NADH pools in strain
NZ9000(pNZpyk) and in strain NZ9000(pNZ8020), harbouring the vector plasmid without the pyk insert. The
kinetics of consumption of [1-13C]glucose (80 mM), the
build-up and depletion of FBP, and the evolution of NAD+
and NADH under an argon atmosphere are shown in Fig. 3.
In starved cells of NZ9000(pNZpyk), the concentration
of NAD+ was approximately 6 mM. The NAD+ pool decreased steadily during glucose utilization, reaching a minimum at about 3?5 mM. At this stage the pool of NADH was
approximately 1?8 mM. Thereafter, NAD+ recovered to its
initial concentration while NADH returned to undetectable
levels. The time-course for the pyridine nucleotide pools
was different in strain NZ9000(pNZ8020). The NAD+ pool
was constant while glucose was available but decreased
sharply at the onset of glucose depletion, reaching the
minimum value of 1?7 mM. Thereafter, NAD+ recovered
very slowly. A similar behaviour has been reported for the
model strain MG1363 (Neves et al., 2002b).
Characterization of pyridine nucleotides kinetics
during anaerobic glycolysis
Quantification of phosphorylated metabolites
during growth
Recently, we described the on-line determination of 13Cglycolytic intermediates, NAD+ and NADH in resting cells
Strains NZ9000(pNZpyk) and NZ9000(pNZ8020) were
grown under anaerobic conditions at pH 6?5 in defined
http://mic.sgmjournals.org
1107
A. Ramos and others
Fig. 3. Kinetics of [1-13C]glucose (80 mM) consumption and build-up/depletion of FBP, NAD+ and NADH pools, as
monitored by in vivo 13C-NMR in strains NZ9000(pNZpyk) (a) and NZ9000(pNZ8020) (b) during glycolysis under anaerobic
conditions. Glucose was added at time 0. X, Glucose; m, FBP; %, NAD+; $, NADH.
medium containing 1 ng nisin ml21. The growth rate of
NZ9000(pNZpyk) (0?42±0?03 h21) was consistently lower
than that of NZ9000(pNZ8020) (0?51±0?02 h21). Ethanol
extracts were prepared from culture samples in the midexponential (OD600 2?2) and stationary (OD600 4?5) growth
phases. Fig. 4 shows the pools of phosphorylated glycolytic
metabolites determined in cell extracts by 31P-NMR. The
profile of glycolytic intermediates in the mid-exponential
phase was similar in both strains examined, the major
metabolites being G6P and FBP. Extracts obtained from
NZ9000(pNZ8020) in the stationary growth phase contained
3-PGA (23 mM), 2-PGA (2?6 mM) and PEP (7?1 mM),
the metabolites typical of starvation in L. lactis, whereas in
Fig. 4. Concentrations of phosphorylated metabolites during
growth. Strains NZ9000(pNZpyk) and NZ9000(pNZ8020) were
grown in defined medium containing 1 % (w/v) glucose,
1 ng nisin ml”1 and 5 mg chloramphenicol ml”1, at pH 6?5 and
30 6C. Ethanol extracts were obtained in the mid-exponential or
stationary growth phases as described in Methods. The data
shown are means of three independent determinations.
1108
NZ9000(pNZpyk) none of these metabolites accumulated
above the detection limit.
Determination of enzyme activities
The activity of LDH, PK, PFK and NADH oxidase was assessed
in extracts obtained from cultures of NZ9000(pNZpyk) and
NZ9000(pNZ8020) growing under either anaerobic or
aerobic atmosphere. The genes pfk, pyk and ldh encoding
PFK, PK and LDH constitute the las operon (Llanos et al.,
1993). Therefore, to assess whether the utilization of the
nisin transcription induction system affects indirectly expression of the las operon, LDH, PK and PFK activities were
also measured in extracts of the parental strain NZ9000
(Table 1). The activity of NADH oxidase was similar in
NZ9000(pNZpyk) and NZ9000(pNZ8020) and, as expected,
it was induced by growth in the presence of oxygen.
However, the activity of LDH was significantly lower in
the PK-overproducing strain. The LDH specific activity
decreased by about 50 % in NZ9000(pNZpyk) grown under
anaerobic conditions when compared to NZ9000(pNZ8020)
or NZ9000, which exhibited similar levels of LDH. This
effect was even more pronounced in extracts of the overproducer strain grown under an oxygen atmosphere. The
activity of PFK was reduced by approximately 50 % in extracts
of the two strains carrying plasmids [NZ9000(pNZpyk) and
NZ9000(pNZ8020)] when compared to NZ9000.
The effect of several effectors, such as Pi, NAD+, NADH, 3PGA and ATP, on the activity of PK (pure enzyme) was
assessed by 31P-NMR. The activity was not affected by
NADH or NAD+, added at concentrations of 2 or 5 mM,
respectively; however, ATP and 3-PGA (added at a concentration of 10 mM) caused a reduction in PK activity
to 30 and 60 %, respectively (data not shown). Under the
experimental conditions used here, the concentration of
Pi required to give 50 % inhibition was 1?5 mM, a value
comparable to that determined for the lactococcal enzyme
Microbiology 150
Overproduction of pyruvate kinase in L. lactis
Table 1. Enzyme activities in extracts of L. lactis NZ9000 and derivative strains grown under
anaerobic or aerobic conditions
Assays were performed under non-controlled atmosphere. Enzyme activities are reported as units (mg
protein)21±standard deviation (n¢4). ND, Not determined.
Enzyme
PFK
PK
LDH
NADH oxidase
NZ9000(pNZpyk)
NZ9000(pNZ8020)
NZ9000
Argon
Oxygen
Argon
Oxygen
Argon
0?87±0?11
27?6±0?8
15?7±0?4
0?07±0?0
ND
0?85±0?10
1?8±0?1
28?0±0?6
0?05±0?0
ND
25?2±0?7
3?8±0?2
0?14±0?01
1?5±0?1
11?6±1?0
0?15±0?02
1?5±0?2
1?5±0?2
31?9±1?3
with the conventional NADH-coupled assay (Crow &
Pritchard, 1976).
DISCUSSION
The complexity of metabolic and regulatory networks is a
great challenge to an integrated view of how individual
components contribute to the overall cell function even in
an organism with the apparent metabolic simplicity of L.
lactis. NMR spectroscopy is a useful methodology since
metabolic changes can be monitored in real time under
physiological conditions. Moreover, as in vivo NMR experiments are usually performed with resting cells, anabolism is
highly depressed, and the metabolic network simplified to
a great extent. Recently, 13C labelling coupled to NMR
detection provided a description of the dynamic behaviour
of glucose, end products, pools of glycolytic intermediates,
NAD+ and NADH, and showed that the pool of reduced
pyridine nucleotides is not a primary factor in the regulation of the glycolytic flux in resting cells of L. lactis (Neves
et al., 2002b).
The present work takes advantage of a similar experimental
approach to provide a detailed picture of glycolysis in a
strain overproducing PK, a key glycolytic enzyme assumed
to play a major role in the regulation of glycolysis. The most
notable consequences of the overproduction of PK were
the acceleration of the rates of FBP depletion and NAD+
recovery once glucose was exhausted, and the lack of
accumulation of 3-PGA and PEP, the two metabolites
associated with cell starvation. In strain NZ9000(pNZ8020),
the level of FBP decreased very slowly after reaching the
intracellular level of about 20 mM, a profile identical to that
of strain MG1363, the parent strain of NZ9000 (Neves et al.,
2002b). This observation denotes a constriction at the level
of PK, further evidenced by the accumulation of metabolites
immediately upstream of pyruvate, namely 3-PGA and PEP
(Neves et al., 2002a). The higher rates of FBP depletion and
of NAD+ recovery, and the absence of detectable pools of
3-PGA or PEP in the PK-overproducing strain, suggest the
occurrence of a metabolic bottleneck at the level of PK in
NZ9000(pNZ8020) when glucose became limiting; in the
PK overproducer this constriction was no longer present,
http://mic.sgmjournals.org
ND
allowing, in addition to a rapid FBP consumption, a complete carbon flow to lactate, thus preventing accumulation
of 3-PGA and PEP. The obstruction at the level of PK is
primarily ascribed to accumulation of Pi, a well-known
inhibitor of the enzyme, counteracting the activating
properties of FBP (Crow & Pritchard, 1976, 1982). In fact,
when glucose transport stopped, the concentration of Pi rose
abruptly due to the release of phosphate engaged mainly in
the large FBP pool and other phosphorylated metabolites
(Fig. 1). The data provide clear evidence for the role of Pi
as an important regulator of the flux through PK in vivo.
The absence of 3-PGA and PEP (below the detection limit of
the technique) is even more striking when aerobic conditions are considered (Fig. 2c) since this and earlier studies
have shown that the levels of these metabolites are greatly
increased in the presence of oxygen (Fig. 1b versus Fig. 2d
and Neves et al., 2002a). In strain NZ9000(pNZ8020)
approximately 50 % of the FBP pool was converted to 3PGA and PEP (Fig. 2d), meaning that the amount of carbon processed via PK, once glucose was depleted, was low.
However, this bottleneck at the level of PK was completely
overcome by 15-fold overproduction of PK (Fig. 2c).
It has been shown that the NAD+ pool is maintained at the
maximum level in MG1363 when glucose is actively metabolized (Neves et al., 2002b). Surprisingly, the level of
NAD+ decreased gradually during glucose consumption in
the PK overproducer, indicating that NADH reoxidation
was less efficient in this strain. We think that this is caused
by the depression of the LDH activity in NZ9000(pNZpyk),
an unexpected lateral effect of the genetic manipulation
strategy pursued here. One could speculate that the nisin
induction system exerts a deleterious effect on the general
transcription machinery and indirectly affects expression of
the pfk–pyk–ldh (las) operon. However, this explanation is
not satisfactory since, at least under anaerobic conditions,
the PK and LDH activities of NZ9000 were not appreciably
affected by the presence of the vector plasmid. In addition,
the low levels of LDH in strain NZ9000(pNZpyk) could not
be due to downregulation of transcription of the las operon
mediated by PK overexpression, since the levels of PFK were
similar in both engineered strains used in the present work
(Table 1). Surprisingly, Andersen et al. (2001a) observed
1109
A. Ramos and others
that alterations of the las promoter affect differentially the
expression of the three genes of the operon. One explanation
could be that the las operon is subjected to posttranscriptional regulation, which results in alterations of its mRNA
fate and consequently modulates the translation efficiency
of its genes. This view is supported by the observation of
processed species of the las mRNA (Luesink et al., 1998).
If this is the case, the presence of chloramphenicol (a
translation inhibitor) during growth of cultures carrying
plasmids could differentially affect expression of pfk and
ldh. Moreover, the pronounced inhibitory effect in the
synthesis of LDH could explain the lower growth rate of
NZ9000(pNZpyk) than that of NZ9000(pNZ8020).
The level of LDH in NZ9000(pNZpyk) was reduced to
approximately 50 % of that of NZ9000, and this may be
sufficient to create a bottleneck at the level of LDH and
hinder the regeneration of NAD+. In fact, according to
Andersen et al. (2001b), L. lactis has limited excess capacity
of LDH, only 70 % more than needed to sustain the lactate
flux in the wild-type cells. Interestingly, despite the lower
NAD+/NADH ratio observed in the PK-overproducing
strain, glucose was metabolized largely to lactate as in the
wild-type strain. Therefore, it seems that the redistribution
of carbon flux in the pyruvate node does not depend
critically on the level of NAD+, and the NAD+/NADH ratio
can vary to some extent without causing noticeable changes
in the composition of end products.
The decreased LDH activity was probably also the main
reason for the shift towards acetate and acetoin production,
under aerobic conditions. The production of acetate and
acetoin was considerably higher in the PK-overproducer
than in NZ9000(pNZ8020), since the activity of LDH was
approximately threefold lower in the former strain while
NADH oxidase activity was similar in both strains.
Overexpression of the gene encoding PK, a postulated
regulatory enzyme, did not lead to an increased glycolytic
flux in non-growing cells, instead a slight decrease was
observed; also, the growth rate of the PK-overproducer was
reduced (17 %). It is worth mentioning that the patterns of
glycolytic intermediates in growing cells (Fig. 4) were
similar to those measured by in vivo NMR in non-growing
cell suspensions, namely the absence of 3-PGA in the PKoverproducer after glucose depletion. Lower growth and
glucose consumption rates in this mutant may be a consequence of (i) diminished level of LDH, and/or (ii) protein
burden, as observed in Zymomonas mobilis (Snoep et al.,
1995) and E. coli, where overproduction of PK or PFK led to
reduced glucose bioconversion rates relative to the control
(Emmerling et al., 1999).
A recent study using metabolic control analysis excluded a
role in the control of flux for the enzymes encoded by the
las operon in growing cells of L. lactis MG1363 (Koebmann
et al., 2002b). The same laboratory has reported that the
demand for ATP exerts a strong control on the glycolytic
flux of resting cells (Koebmann et al., 2002a). We found that
1110
the glycolytic flux in resting cells of L. lactis did not increase
in response to a 15-fold increase in the level of PK, but
when glucose became limiting, PK overproduction allowed
a greater flux through PK and led to complete channelling
of carbon to end products without accumulation of the
glycolytic intermediates typical of starvation.
ACKNOWLEDGEMENTS
This work was supported by the BIOTECH Programme, contract
QLK1-CT-2002-02388 of the Commission of the European Community and by Fundação para a Ciência e Tecnologia (FCT), Portugal,
contract PRAXIS/P/BIA/11072/98 and FEDER. The work at the Centro
de Investigaciones Biológicas was partially supported by the Program
of Strategic Groups of the Comunidad de Madrid. A. Ramos
acknowledges a fellowship of FCT. We thank A. Mingote for the
purification of PK and NIZO Food Research for supplying L. lactis
subsp. cremoris NZ9000 and plasmid pNZ8020.
REFERENCES
Andersen, H. W., Solem, C., Hammer, K. & Jensen, P. R. (2001a).
Twofold reduction of phosphofructokinase activity in Lactococcus
lactis results in strong decrease in growth rate and glycolytic flux.
J Bacteriol 183, 3458–3467.
Andersen, H. W., Pedersen, M. B., Hammer, K. & Jensen, P. R.
(2001b). Lactate dehydrogenase has no control on lactate production
but has a strong negative control on formate production in
Lactococcus lactis. Eur J Biochem 268, 6379–6389.
Bailey, J. E. (2001). Reflections on the scope and the future of
metabolic engineering and its connections to functional genomics
and drug discovery. Metab Eng 3, 111–114.
Bradford, M. M. (1976). A rapid and sensitive method for quantification of microgram quantities of protein utilizing the principle of
protein-dye binding. Anal Biochem 72, 248–254.
Crow, V. L. & Pritchard, G. G. (1976). Purification and properties of
pyruvate kinase from Streptococcus lactis. Biochim Biophys Acta 438,
90–101.
Crow, V. L. & Pritchard, G. G. (1982). Pyruvate kinase from
Streptococcus lactis. Methods Enzymol 90, 165–170.
de Ruyter, P. G., Kuipers, O. P. & de Vos, W. M. (1996). Controlled
gene expression systems for Lactococcus lactis with the food-grade
inducer nisin. Appl Environ Microbiol 62, 3662–3667.
Dornan, S. & Collins, M. A. (1990). High efficiency electroporation of
Lactococcus lactis subsp. lactis LM0230. Lett Appl Microbiol 11, 62–64.
Emmerling, M., Bailey, J. E. & Sauer, U. (1999). Glucose catabolism
of Escherichia coli strains with increased activity and altered regulation of key glycolytic enzymes. Metab Eng 1, 117–127.
Fordyce, A. M., Moore, C. H. & Pritchard, G. G. (1982). Phospho-
fructokinase from Streptococcus lactis. Methods Enzymol 90, 77–82.
Garrigues, C., Loubiere, P., Lindley, N. D. & Cocaign-Bousquet, M.
(1997). Control of the shift from homolactic acid to mixed-acid
fermentation in Lactococcus lactis: predominant role of the NADH/
NAD+ ratio. J Bacteriol 179, 5282–5287.
Gasson, M. J. (1983). Plasmid complements of Streptococcus lactis
NCDO 712 and other lactic streptococci after protoplast-induced
curing. J Bacteriol 154, 1–9.
Koebmann, B. J., Solem, C., Pedersen, M. B., Nilsson, D. & Jensen,
P. R. (2002a). Expression of genes encoding F1-ATPase results in
Microbiology 150
Overproduction of pyruvate kinase in L. lactis
uncoupling of glycolysis from biomass production in Lactococcus
lactis. Appl Environ Microbiol 68, 4274–4282.
Koebmann, B. J., Andersen, H. W., Solem, C. & Jensen, P. R.
(2002b). Experimental determination of control of glycolysis in
Lactococcus lactis. Antonie van Leeuwenhoek 82, 237–248.
Kuipers, O. P., De Ruyter, P. G. G. A., Kleerebezem, M. & de Vos,
W. M. (1998). Quorum sensing-controlled gene expression in lactic
acid bacteria. J Biotechnol 64, 15–21.
Leenhouts, K. J., Kok, J. & Venema, G. (1989). Campbell-like
integration of heterologous plasmid DNA into the chromosome of
Lactococcus lactis subsp. lactis. Appl Environ Microbiol 55, 394–400.
Llanos, R. M., Harris, C. J., Hillier, A. J. & Davidson, B. E. (1993).
O’Sullivan, D. J. & Klaenhammer, T. D. (1993). Rapid mini-prep
isolation of high-quality plasmid DNA from Lactococcus and
Lactobacillus spp. Appl Environ Microbiol 59, 2730–2733.
Poolman, B. & Konings, W. N. (1988). Relation of growth of
Streptococcus lactis and Streptococcus cremoris to amino acid
transport. J Bacteriol 170, 700–707.
Poolman, B., Smid, E. J., Veldkamp, H. & Konings, W. N. (1987).
Bioenergetic consequences of lactose starvation for continuously
cultured Streptococcus cremoris. J Bacteriol 169, 1460–1468.
Ramos, A. & Santos, H. (1996). Citrate and sugar cofermentation in
Leuconostoc oenos, a 13C nuclear magnetic resonance study. Appl
Environ Microbiol 62, 2577–2585.
Identification of a novel operon in Lactococcus lactis encoding three
enzymes for lactic acid synthesis: phosphofructokinase, pyruvate
kinase, and lactate dehydrogenase. J Bacteriol 175, 2541–2551.
Ramos, A., Boels, I. C., de Vos, W. M. & Santos, H. (2001).
Luesink, E. J., van Herpen, R. E., Grossiord, B. P., Kuipers, O. P. &
de Vos, W. M. (1998). Transcriptional activation of the glycolytic las
Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning:
operon and catabolite repression of the gal operon in Lactococcus
lactis are mediated by the catabolite control protein CcpA. Mol
Microbiol 30, 789–798.
Mason, P. W., Carbone, D. P., Cushman, R. A. & Waggoner, A. S.
(1981). The importance of inorganic phosphate in regulation
of energy metabolism of Streptococcus lactis. J Biol Chem 256,
1861–1866.
Neves, A. R., Ramos, A., Nunes, M. C., Kleerebezem, M.,
Hugenholtz, J., de Vos, W. M., Almeida, J. S. & Santos, H. (1999).
In vivo NMR studies of glycolytic kinetics in Lactococcus lactis.
Biotechnol Bioeng 64, 200–212.
Neves, A. R., Ramos, A., Costa, H., van Swam, I. I., Hugenholtz, J.,
Kleerebezem, M., de Vos, W. M. & Santos, H. (2002a). Effect of
different NADH oxidase levels on glucose metabolism of Lactococcus
lactis: kinetics of intracellular metabolite pools by in vivo NMR. Appl
Environ Microbiol 68, 6332–6342.
Neves, A. R., Ventura, R., Mansour, N., Shearman, C., Gasson, M. J.,
Maycock, C., Ramos, A. & Santos, H. (2002b). Is the glycolytic flux
in Lactococcus lactis primarily controlled by the redox charge?
Kinetics of NAD+ and NADH pools determined in vivo by 13CNMR. J Biol Chem 277, 28088–28098.
http://mic.sgmjournals.org
Regulation of carbon flux from glycolysis to exopolysaccharide
biosynthesis in Lactococcus lactis. Appl Environ Microbiol 67, 33–41.
a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold
Spring Harbor Laboratory.
Snoep, J. L., Yomano, L. P., Westerhoff, H. V. & Ingram, L. (1995).
Protein burden in Zymomonas mobilis: negative flux and growth
control due to overproduction of glycolytic enzymes. Microbiology
141, 2329–2337.
Solem, C., Koebmann, B. J. & Jensen, P. R. (2003). Glyceraldehyde-
3-phosphate dehydrogenase has no control over glycolytic flux in
Lactococcus lactis MG1363. J Bacteriol 185, 1564–1571.
Thompson, J. (1987). Regulation of sugar transport and metabolism
in lactic acid bacteria. FEMS Microbiol Rev 46, 221–231.
Thompson, J. & Thomas, T. D. (1977). Phosphoenolpyruvate and 2-
phosphoglycerate: endogenous energy source(s) for sugar accumulation by starved cells of Streptococcus lactis. J Bacteriol 130, 583–595.
31
P nuclear magnetic
resonance spectroscopy and 14C fluorography in studies of glycolysis
and regulation of pyruvate kinase in Streptococcus lactis. J Bacteriol
158, 791–800.
Thompson, J. & Torchia, D. A. (1984). Use of
Valentini, G., Chiarelli, L., Fortin, R., Speranza, M. L., Galizzi, A. &
Mattevi, A. (2000). The allosteric regulation of pyruvate kinase. A
site-directed mutagenesis study. J Biol Chem 275, 18145–18152.
1111